Each model run should print or log a wave of useful summary stats

[crahal]

Examples:

• % coefficients with the same sign (e.g. Young and Holsteen style).
• Min and max (and mean?) value of coefficient of interest across all bootstraps and specs.
• The "Best" model selected by user-specified IC

[dhvalden]

New feature: Populate the summary function?

[vallerrr]

I edited the summary() function in the OLSRobut class and the current output looks like this ⬇️

code can be found in the latest commit in the relevant branch and we @crahal @dhvalden should think of any other model summary information we want to add to the current function/ format of the print content before merging the branch into main

[crahal]

This is absolutely great! Maybe we can discuss, but perhaps also:

• Mean (and max\min as appropriate) value of OOS evaluation metric, and the spec which achieves the min\max?
• Model-spec which minimizes the relevant IC
• Something to do with the BMA (although unsure exactly what: the BMA of the beta of interest, perhaps)?

(I don't think we need negative portion of beta; isn't this just 1-positive portion?) (Maybe we want to split the 'robustness metrics' into 'inference' and 'oos'?)

[vallerrr]

updated summary function